1642 - Risk Factors for Major Adverse Cardiovascular Events in Rheumatoid Arthritis Patients Treated with Tocilizumab

Monday, November 7, 2011: 2:45 PM
W375c (McCormick Place West)
Vijay Rao1, Andrey Pavlov2, Micki Klearman1, David Musselman3, Jon T. Giles4, Joan M. Bathon4, Naveed Sattar5 and Janet S. Lee3, 1Genentech, San Francisco, CA, 2Everest, Toronto, 3Roche, Nutley, NJ, 4Columbia University Medical Center, New York, NY, 5University of Glasgow, Glasgow, United Kingdom
Presentation Number: 1642

Background/Purpose: To examine the associations of lipids, inflammation, and rheumatoid arthritis (RA) disease activity measures with risk for cardiovascular events in moderate to severe RA patients treated with tocilizumab (TCZ).

Methods: Post-hoc analyses of the combined data from 5 pivotal TCZ phase-III trials and long-term extension periods were conducted in a stepwise fashion. First, demographic, laboratory, and disease characteristics were compared for patients with (n = 42) and without (n = 3944) major adverse cardiovascular events (MACE: non-fatal myocardial infarction, non-fatal stroke, or death due to cardiovascular cause) determined by independent, blinded adjudication. Univariate Cox proportional hazard modeling was performed to evaluate potential risk factors for MACE at baseline (BL) and post-BL (wk 24) with time to MACE as the time variable. Multivariate Cox proportional hazard models were also fit to obtain best prediction of time to MACE. Statistically significant predictors were considered as those with P<0.05.

Results: For all MACE events occurring during TCZ exposure including open-label extension treatment, univariate modeling revealed BL age, history of coronary artery disease (CAD), disease activity (DAS28), total cholesterol/HDL ratio, and albumin as predictive characteristics (P<.05), but only age and history of CAD consistently remained predictive in a series of multivariate sensitivity analyses. Neither reductions in measurements of inflammation (CRP and ESR), nor increases in lipid parameters at 24 wks were associated with MACE on TCZ therapy. Additional Cox models employing single predictors and adjusting for BL age revealed that the following wk 24 outcomes were statistically significant predictors of future MACE: DAS28 based scores (DAS28, AUC of DAS28, EULAR response), TJC, SJC, and Patient Global Assessment (Table). Additionally, greater reductions in DAS28 score from BL to wk 24 were inversely associated with MACE, after adjusting for age and BL DAS28 score (Table).


Table. Univariate Associations Between RA Parameters and MACE, Adjusted for Age at Baseline

Disease Activity Assessments at Week 24


Hazard Ratio (95% CI)

P Value

# Swollen Joints (28)*

1.101 (1.055, 1.149)


DAS28 score

1.482 (1.201, 1.828)


# Tender Joints (28)

1.069 (1.030, 1.110)


EULAR (Good vs. No Response)

0.219 (0.084, 0.569)


Reduction from baseline in DAS28

0.704 (0.556, 0.891)


AUC of DAS28 to week 24, score-years

2.154 (1.208, 3.840)


Patient Global Assessment Score, mm

1.015 (1.002, 1.028)


*Interpretation of HR: After adjustment for age, one additional SJC at week 24 is associated with an increased 10.1% hazard for MACE.

Observed values; LOCF for missing SJC or TJC at Week 24

Also adjusted for baseline DAS28 score.


Conclusions: Traditional cardiovascular risk factors at BL (history of CAD, total cholesterol/HDL ratio, and albumin), as well as disease activity parameters are associated with MACE for patients on TCZ therapy. Only age and CAD history remained statistically significantly associated with MACE on multivariate analysis of BL characteristics. Further, in this on-treatment analysis, risk of MACE was broadly linked to elevated disease activity at BL and a less robust therapeutic response at week 24. There was no association between lipid change and risk for MACE. We hypothesize that mitigating cardiovascular risk in RA patients may require a multifaceted approach that includes effective control of RA disease activity.

Role of study sponsor: This study was funded by Roche. Support for third-party writing assistance for this abstract was provided by F. Hoffmann-La Roche Ltd.


Keywords: rheumatoid arthritis (RA) and tocilizumab

Disclosure: V. Rao, Genentech, Inc., 3 ; A. Pavlov, Everest, 3 ; M. Klearman, Genentech, Inc., 3 ; D. Musselman, F. Hoffmann-La Roche Ltd, 3 ; J. T. Giles, Fa Hoffmann-La Roche Ltd, 5, Genentech, Inc., 5, IDEC, 5 ; J. M. Bathon, None; N. Sattar, F. Hoffmann-La Roche Ltd, 5, F. Hoffmann-La Roche Ltd, 8 ; J. S. Lee, F. Hoffmann-La Roche, 3 .